Discover and read the best of Twitter Threads about #ML2
Most recents (2)
#ML2 (psyarxiv.com/9654g) has got me thinking about the following thesis: Over the past ~20 odd years (or longer), social psychologists have grossly underestimated how hard it is for people to change
Let me state up front that this thesis is not new. I'm also not sure that social psychologists actually *believe* that people are easy to change. Nevertheless, the assumption of changeability underlies a lot of social psychology research from the past ~20-30 years
The assumption of changeability is most extreme in the various flavors of social priming studies. Being outside a box enhances creativity (journals.sagepub.com/doi/10.1177/09…), exposure to money increases conservatism (ncbi.nlm.nih.gov/pubmed/22774789) and on and on and on
As a #ML2 author, very excited to see this paper finally out! The headline that 50% of effects replicate buries the lead a bit. For me, the major finding here concerns heterogeneity. What's heterogeneity, you ask? Gather round...
Heterogeneity describes how much effects being replicated vary from lab to lab, *beyond* sampling error. Sampling error says we expect that by chance, some labs overestimate the true effect size, others underestimate. Heterogeneity says how much more we expect effects to vary.
If differences between labs are to explain differences in replicability, there MUST be sizable heterogeneity in effects. The magnitude of the between site differences limits how much variance heterogeneity can explain.
